JPH0482178B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION A. Field of Industrial Application This invention relates generally to semiconductor processing, and more specifically to improvements in photoresist technology for reactive ion etching in semiconductor processing applications.

B. Prior art The current reactive method using a single-layer resist system
Ion etching techniques have two drawbacks. First, the photoresist layer cannot be reprocessed after development, and second, the etched metal line width varies due to reflections from detailed structures. FIGS. 12 to 17 illustrate a problem with the prior art in that line widths vary due to reflection from detailed structures. 1st
In Figures 2 through 17, certain conventions have been adopted to indicate the relative orientation of the figures. Figure 12 shows
Contains three rectangular coordinates x, y, and z. 12th
The figures are cross-sectional views as seen from the x-y plane, FIGS. 13 to 16 are cross-sectional views as seen from the y-z plane, and FIG. 17 is a plan view as seen from the x-y plane. 1st
Figure 2 shows a starting composite with a silicon dioxide layer 22, an aluminum copper alloy layer 24, a polycrystalline silicon (polysilicon) layer 26 on top of a substrate 20, with a positive photoresist layer 28 deposited on top. It shows. The silicon dioxide layer 22 has a region A
The provision of a thinner region C, so that in order to join region A with region C, the sloped step region B must become a transition region. In areas such as graded step region B, the light used to expose the photoresist layer 28 produces an undesirable diffraction pattern that causes the width of the exposed portion of the resulting photoresist to vary.
This results in undesirable variations in the width of the reactive ion etched metal lines. This can be clearly seen by looking at FIGS. 13 and 14. Figure 13 shows cutting line 1 in Figure 12.
It is a cutaway view of area A cut along A-1A'. From FIG. 13, when patterned light 30 is used to expose photoresist layer 28, portions 28' are exposed and therefore later developed and removed, and portions 28'' are not exposed and after development. Comparing FIG. 13 with FIG. 14, FIG. 14 shows a cross-sectional view of the sloped step region B of the composite shown in FIG. The light 30 passes through the photoresist layer 28 and is reflected by the polysilicon layer 26 and the aluminum-copper alloy layer 24, and is reflected by the polysilicon layer 26 in region B.
Due to the inclined plane of the diffraction pattern 32 occurs,
Light is directed into the central portion 28'' of the photoresist layer 28, thereby causing unwanted exposure along the edges of the region 28''. FIG. 15 illustrates the next step in the processing of prior art photoresist layer 28 by developing photoresist layer 28 using a developer such as potassium hydroxide. The photoresist layer 28 may be, for example, a Novolak type photoresist. The photoresist area 28'' in area A is shown to be wider than the photoresist area 28'' in area B after development. As mentioned above, this is due to the light diffraction pattern that occurs within region B, and this diffraction pattern causes the developer to
Melt the 8″ addition and narrow it down.
FIG. 16 shows a prior art processing technique in which a photoresist pattern is used as a mask to perform reactive ion etching on polysilicon layer 26 and aluminum-copper alloy layer 24. Indicates the next step. As can be seen from FIG. 16, the line widths of the resulting polysilicon layer 26 and aluminum-copper alloy layer 24 are narrower in region B than in region A. This is easy to understand when looking at the plan view in Figure 17. In Figure 17,
The line width of the polysilicon layer is wider in region A than in region B. This is an undesirable feature of this prior art and is a result of the uncontrollable light 32 diffracted from detailed structures such as occurs within region B. This prior art technique involves adding a light-absorbing dye to the photoresist layer 28 to create a polysilicon layer 26.
and by reducing the amount of light reflected from the aluminum-copper alloy layer 24 to the photoresist layer 28.
We are trying to correct this problem.

C. Problems to be Solved by the Invention The problem with this prior art method is that in order for the light 30 coming from above to penetrate sufficiently to the bottom of the photoresist layer 28, the intensity of the light must be very high. As a result, the photoactivity of the photoresist decreases. The presence of a light-absorbing dye in the photoresist layer 28 creates a large gradient in exposure between the top and bottom of the photoresist, which creates problems with the vertical contours of the resulting metal structures etched after the photoresist is developed.

Another problem with the prior art is that the photoresist structure cannot be reworked once it has been exposed and developed. This can be seen by looking at the prior art diagrams of FIGS. 18-20. In FIG. 18, the photoresist layer 28 is shown as described above.
is deposited on top of the polysilicon layer 26,
This photoresist layer 28 is exposed to patterned light 30 as described above. As a result, the photoresist is exposed in area 28' and subsequently removed in a development step. The central region 28'' is not exposed to light and is not removed after the development step. FIG. 19 shows that potassium hydroxide, sodium metasilicate, etc., react with the portions 28' of the photoresist layer exposed to the light 30 and remove it. The development step is shown using a conventional developer which dissolves the unexposed portion 28'', but does not dissolve the unexposed portion 28''.
A problem with the development step is that the developer, such as potassium hydroxide, will also significantly etch the exposed surfaces of the polysilicon layer 26. The etched surface 34 shown in FIGS. 19 and 20 remains as a permanent feature on the surface of the polysilicon layer 26 after the development step. Examination of the remaining photoresist structure 28'' using an optical microscope reveals that the photoresist structure 28'' is similar to other structures on the integrated circuit chip.
The development step must be performed in such a way that it can be determined whether the 8" This misalignment measurement is very important as other disturbances will occur.If it is determined that there is a misalignment, the photoresist structure 2
A rework cycle must be performed in which 8' is removed from the surface of polysilicon layer 26, a new photoresist layer is deposited, and exposed to patterned light. Referring to FIG. 20, the erroneously etched portion 34 of the polysilicon layer 26 remains intact and presents a surface that is difficult to rework due to its misleading reflectivity when inspected in a subsequent alignment measurement step. I understand the problem. Typically, integrated circuit wafers whose photoresist structures are determined to be misaligned after photoresist processing are discarded because the latent image is etched into the polysilicon layer 26 by the developer. There must be.

Accordingly, it is an object of the present invention to provide an improved photoresist process for providing metal patterns in semiconductor devices by reactive ion etching.

It is another object of the present invention to provide an improved photoresist process that minimizes back reflections and diffraction of photoresist exposure step light.

It is another object of the present invention to provide an improved photoresist process that is reworkable.

D. Means for Solving the Problems The above and other objects, features and advantages of the present invention are realized by the photoresist process disclosed herein. This method is based on applying a layer of release agent. The release agent is a solution of a polysulfone composition and a dye that is applied to the surface of the polysilicon layer. A photoresist layer can then be deposited on top of this release layer.
The photoresist is then exposed, developed, and its alignment measured. If the resulting photoresist structure proves to be incorrectly aligned, reprocessing can be easily carried out by dissolving the stripping agent in a suitable solvent. This allows existing photoresist structures to be removed without leaving latent images in the polysilicon layer. A new stripper and dye solution can then be applied, followed by a new layer of photoresist and the rework cycle can continue.
A new photoresist layer can then be exposed with an exposure pattern, developed and its alignment measured once again. If the original photoresist layer or reworked photoresist layer is found to be properly aligned, normal processing steps can proceed. Typical processing steps include plasma hardening of the photoresist, followed by reactive ion etching of the release layer, and then firing of the photoresist. Thereafter, reactive ion etching can be performed on the polysilicon layer and the aluminum-copper alloy layer, which are masked by the release layer and photoresist composite. The resulting etched polysilicon and aluminum-copper structure can then be processed by dissolving the remaining release layer and removing the remaining photoresist structure. This method also provides improved control over back-reflected and diffracted light during the exposure step. Because the dye is selected to have an approximately maximum absorption coefficient with respect to the wavelength of maximum energy of the exposing light, no portion of the light that passes through the photoresist and enters the release layer is reflected or diffracted. It can be absorbed before it has a chance to return to the photoresist layer. This allows for uniform exposure of the photoresist even when passing over structures that underlie the release layer and may otherwise cause undesirable light reflection or diffraction. Furthermore, due to the absence of such dyes in the photoresist layer, exposing light is transmitted more uniformly through the photoresist layer, resulting in uniform exposure of the photoresist.

E Example Figure 1 shows the steps of the etching method of the present invention, and Figures 2 to 1 show the structures in successive steps.
The etching method of the present invention will be explained in detail with reference to FIG. Substrate 20 shown in FIG.
The process flow begins at step 50 with the composite of polysilicon layer 26, aluminum-copper alloy layer 24, and silicon dioxide layer 22 on top of the polysilicon layer 26, and includes a third step of applying a release layer 40 to the surface of polysilicon layer 26. Proceed to figure. The release agent is referred to herein as polysulfone. Its specific composition is as follows. The intermediate solution consists of a mixture of 45.75% by weight of a solvent such as N-methylpyrrolidone and 45.75% by weight of a diluent such as diethylene glycol dimethyl ether. To this solution is added 8.5% by weight of a release agent such as 5003P-polyethersulfone powder. This powder can be used, for example, from ICI Chemical (ICI Chemical).
Company). Dissolve this powder in the liquid ingredient. The solvent component is 35 to 55% by weight, the diluent component is 35 to 55% by weight, and the powder component is 4 to 55% by weight.
It can be 15% by weight. Then this intermediate solution
To the 98% by weight is added 2% by weight of a suitable light-absorbing dye having approximately the maximum absorption coefficient (exhibiting an absorption peak) with respect to the wavelength of maximum energy of the light used to expose the photoresist. For example, a projection means such as a GCA4800DSW step projector with a maximum radiated energy wavelength of 436 nm can be used. A suitable light-absorbing dye for this wavelength is, for example, Orosol Yellow 4GN monoazo dye, commercially available from Chiba-Geigy Corporation. The dye component is 0.5 to 5 when mixed with the intermediate solution.
It can be expressed as % by weight. When the resulting solution is stirred and allowed to stand for several hours, it becomes ready to be applied as the polysulfone release layer 40 shown in FIG. The coating thickness of polysulfone release layer 40 is suitably about 0.3 microns. A typical application condition is, for example, a spin-on application step.

The process then proceeds to step 52 of depositing a photoresist layer 28 on the polysulfone release layer 40, as shown in FIG. Suitable positive photoresist materials are of the Novorak resin type, which are conventionally known in the art.

The next step 54 is to expose photoresist layer 28 with a light pattern as shown in FIG. The resulting exposed photoresist layer is developed with a suitable developer such as 0.2N potassium hydroxide. This developer dissolves the portions 28' of the photoresist layer that were exposed to light 30 in the exposure step. The resulting photosulfone structure 28' is shown on top of the polysulfone layer 40 in FIG.

At this point, step 58 of FIG. 1 can be entered to measure the alignment of photoresist structure 28'' with respect to other structures on the semiconductor chip.
If the alignment is found to be incorrect at step 60 of FIG. 1, a rework cycle is shown beginning at step 62.

In the rework cycle, release layer 40 is dissolved using N-methylpyrrolidone as a suitable solvent. The photoresist structure 28'' is then freed and floats for easy removal. The process then returns to step 50 of FIG. 1 and as shown in FIG.
followed by the deposition of a new photoresist layer 28A in step 52. The process of FIG. 1 then proceeds to steps 54 and 56 in which the developed photoresist structure reappears and proper alignment can be determined in step 58.

If the alignment of the photoresist structure 28'' is determined to be good at step 60, the process proceeds to a normal processing cycle at step 64. Step 64 is a plasma curing step of the photoresist structure 28'' in a plasma etching chamber. The photoresist structure 28'' is cured using a suitable plasma material, such as a fluorocarbon, in preparation for the next reactive ion etching step. The process then proceeds to a conventional process step 66 to reactively ion etch the release layer 40. A suitable plasma etching composition is oxygen plasma, which provides rapid reactive ion etching.
The step removes the polysulfone release layer 40 covered by the photoresist structure 28'', exposing the polysilicon layer 26 (FIG. 9).
A hardening bake step 68 is then performed to further harden the photoresist layer. This curing and baking step can be carried out at 135°C and is suitable for hard photoresists.
A masking structure 28'' is obtained which is used to mask portions of the underlying polysilicon layer and aluminum-copper alloy layer that are not desired to be etched in the next reactive ion etching step. Step 70 then proceeds to reactive ion etch the polysilicon layer 26 onto the aluminum copper alloy layer 24 using a chlorinated gas mixture copper suitable reactive ion etch composition well known in the art. The structure at this point is shown in Figure 10. In Figure 1, the last major step 72 in the conventional process is to dissolve the release layer 40 with N-methylpyrrolidone to release the remaining photoresist structure 28''. Let it float,
This allows it to be removed. Thereafter, the wafer can be cleaned and a forming gas anneal step performed to provide the final etched structure of polysilicon layer 26 and aluminum-copper alloy layer 24 shown in FIG. During this anneal step, the polysilicon is alloyed with aluminum-copper.

This method has the advantage of providing uniform linewidths due to increased control over the back-reflected and diffracted rays during exposure of the photoresist. This method also allows for multiple reworking of the photoresist structure due to the presence of a polysulfone release layer that protects the underlying polysilicon layer from being accidentally etched by the developer during the photoresist development step. It has the advantage that it can also be used.

In another embodiment of the invention, if the polysilicon layer or the metal layer or silicon dioxide layer has a large number of vertical protrusions, a modified image reversal photo film is applied to the top surface of the polysilicon layer before applying the polysulfone release layer 40. The inclusion of a layer of planarizing material, such as a resist, is suitable. Furthermore, it is often advantageous to apply an adhesion promoter to the polysilicon layer before applying the planarizing layer or the polysulfone release layer. A suitable adhesion promoter for applying to the surface of polysilicon layer 26 before applying the polysulfone layer or planarization layer is, for example, mexamethylene disilane.

F. Effects of the Invention The photoresist can be uniformly exposed and reprocessed easily.

[Brief explanation of drawings]

FIG. 1 is a flow diagram of the method of the present invention showing the major steps in carrying out the method. FIG. 2 is a schematic cross-sectional view showing the beginning of the processing steps of the invention disclosed herein, in which a silicon dioxide layer 22 is deposited on the top surface, followed by an aluminum-copper alloy layer 22.
4 is shown with a polysilicon layer 26 deposited thereon, followed by a polysilicon layer 26. FIG. FIG. 3 is a schematic cross-sectional view of the first major step of the method of the present invention, which is the application of a polysulfone layer 40 on top of the polysilicon layer 26. FIG. 4 shows that a photoresist layer 28 is applied to the top surface of the polysulfone layer 40.
3 is a schematic cross-sectional view of the next major step of the method of the invention; FIG. FIG. 5 is a schematic cross-section of the next major step in the method of the present invention, in which the photoresist layer 28 is exposed to patterned light 30, the transmitted portion of which is absorbed by the underlying polysulfone layer 40. It is a diagram. FIG. 6 is a schematic cross-sectional view of the next major step of developing the photoresist layer 28, leaving the photoresist structure 28'' on top of the polysulfone layer 40. FIG. illustrates the step after performing alignment measurements and determining that the photoresist structure is not properly aligned, by melting the photoresist structure 28'' and thereby initiating a rework cycle by removing the photoresist structure 28''. It is a schematic sectional view. FIG. 8 is a schematic cross-sectional view showing the continuation of the rework cycle in which a new polysulfone stripper and dye solution layer 40A is applied, followed by a new photoresist layer 28A. next,
The rework cycle continues by exposing the photoresist with the original pattern, developing the photoresist, and taking another alignment measurement. FIG. 9 is a schematic diagram illustrating the continuation of the conventional processing steps of reactive ion etching of the polysulfone release layer 40, followed by plasma curing of the photoresist after the steps shown in FIG. FIG. This step is usually followed by a hardening bake step to harden the photoresist layer. Figure 10 shows
The next major step is the reactive ion etching of polysilicon layer 26 and aluminum-copper alloy layer 24 masked with polysulfone layer 40 and photoresist layer 28''. Figure 2 is a schematic illustration of the last major step of the method, which is dissolving the mask in a suitable solvent so that the remaining masking photoresist structure 28'' can be removed. FIG. 12 shows a silicon dioxide layer 22 having a thick region A, a thin region C, and a sloped region B, with an aluminum-copper alloy layer 24, a polysulfone layer 26, and a photoresist layer 28 deposited on top thereof. 1 is a cross-sectional view from the xz plane of a prior art semiconductor structure; FIG. FIG. 13 shows the exposure of photoresist layer 28 in area A at section line 1A of FIG.
FIG. 2 is a cross-sectional view taken from the yz plane along -1A'. FIG. 14 is a cross-sectional view taken from the yz plane along section line 1B-1B' of FIG. 1, showing the exposure of the photoresist layer and region B. Figure 15 shows
Areas A and B aligned with each other such that the relative width of the developed photoresist 28'' in area A can be compared with the relative width of the photoresist in area B.
FIG. 2 is a sectional view seen from the yz plane. 1st
Figure 6 is a cross-section in the y-z plane showing the next step of reactive ion etching of the polysilicon layer and the aluminum-copper alloy layer masked by the photoresist layer in areas A and B. FIG. 3 is a diagram in which the relative widths of metal lines obtained in region A and region B can be compared. FIG. 17 is a plan view of the width of the obtained metal line viewed from the xy plane, showing that the width of the metal line is narrower in region B than in region A. FIG. 18 is a schematic cross-sectional diagram illustrating a prior art technique for exposing photoresist layer 28 with a patterned light source. FIG. 19 is a schematic cross-sectional view illustrating the next step of developing the photoresist layer so that the remaining photoresist structure 28 remains intact, depicting the surface 34 where polysilicon layer 26 was accidentally etched. There is. FIG. 20 shows the photoresist structure 28'' after it has been determined that the photoresist structure 28'' is misaligned and that there is an undesirable latent image 34 which is the etched surface of the polysilicon layer 26.
FIG. 3 is a schematic cross-sectional view showing the next step in a prior art rework cycle where the rework cycle is removed. 20...Substrate, 22...Silicon dioxide layer, 24
... Aluminum copper alloy layer, 26 ... Polycrystalline silicon layer, 28 ... Photoresist layer, 30 ...
Light, 40...Polysulfone release layer.

Claims (1)

[Claims] 1. A first step of forming a release layer by depositing a solution containing a release agent and a dye on the surface of the object to be etched; and a second step of forming a photoresist layer on the release layer. a third step of exposing the photoresist layer to a pattern of light; a fourth step of developing the photoresist layer; examining the alignment of the remaining photoresist layer after the development; If the alignment is good, proceed to the seventh step below; if the alignment is poor, proceed to the sixth step below; a fifth step, in which the release layer is dissolved and removed together with the photoresist layer; a sixth step of redoing the first to fifth steps; a seventh step of performing reactive ion etching on the peeling layer using the remaining photoresist layer as a mask; an eighth step of performing reactive ion etching on the surface to be etched using the layer and the remaining release layer thereunder as a mask; and a step of dissolving the remaining release layer and removing it together with the remaining photoresist layer. 9. A reactive ion etching method comprising step 9. 2 4-15% by weight of polyether sulfone powder, 35-55% by weight of N-methylpyrrolidone, 35
A dye having approximately the maximum absorption coefficient with respect to the wavelength of the maximum radiant energy for photoresist exposure is added in an intermediate solution consisting of ~55% by weight of diluent such that the proportion of the total solution is 0.5 to 5% by weight. A solution for forming a release layer, which is produced by adding to the solution and used for forming a release layer in a reactive ion etching method.